Mesh aggregation is implemented as an option in the process settings for 3D Warp analysis to improve solution speed and reduce memory requirement for 3D analysis of typical thin-walled plastic parts. This option is enabled by default.
Mesh aggregation is a technique which reduces the number of layers of elements in a tetrahedral mesh to two, while upgrading the elements to 2nd-order, 10-node tetrahedral elements. Because the 2nd-order element is a high-quality element, even two layers of 2nd-order tetrahedral elements can provide equally good deflection (warpage) results for thin-walled parts compared to six layers of 2nd-order elements. By contrast, a tetrahedral mesh used for 3D Fill+Pack analysis typically has six or more layers of 1st-order, 4-node tetrahedral elements.
When mesh aggregation is enabled, two different meshes are used for 3D Fill+Pack and 3D Warp analyses, respectively. Using mesh aggregation, there are fewer elements and fewer unknowns in the 2-layer mesh, and therefore, the 3D Warp memory requirement is reduced significantly, and the element-level computations and equation solution time are much less.
If the part is solid or thick (true 3D geometry), using mesh aggregation is not recommended. Turning off the mesh aggregation option causes the original mesh to be used for the analysis. This increases analysis time and memory requirement but improves accuracy for true 3D parts.
When the mesh aggregation option is disabled, you can access additional mesh control options. For example, you can specify whether the 1st-order tetrahedral elements created by the mesher should be upgraded to 2nd-order tetrahedral elements for the 3D Warp analysis (without reducing the number of layers of elements). In thin-walled areas of the part, using 2nd-order elements will improve the accuracy of the prediction. The Upgrade tetrahedral elements to second order option is set to Automatic by default. In this case, 3D Warp automatically uses a scheme in which 1st-order, 4-node tetrahedral elements are used in the true 3D, solid areas, and 2nd-order, 10-node tetrahedral elements are used in the thin-walled areas, with transitional elements connecting these areas.